AVS Forum banner

why don't plasmas have a higher refresh rate?

25K views 129 replies 33 participants last post by  lovingdvd 
#1 ·
Been looking at the new Samsung's and Panasonics in the store, and I see flicker. Especially in bright scenes and large screens and at the edges(peripheral vision). I know from playing around with CRT tubes that I stop seeing flicker at about 75hz and over aprox.85hz I can't even see it if I try by using peripheral vision. My pre Kuro Pioneer has refresh settings, and I have it at 75hz and don't notice any flicker.

Do the newest Plasmas have higher than 50/60hz refresh settings, and if they don't why not? I´m guessing that a higher refresh could cause more image retention, more likely burn in, more dithering and posturizing(kinda like how DLP´s will have more if the color wheel spins faster)?


p.s. I´ve been noticing the flicker here in the USA, so over in Europe(where I live) it's probably gonna be even worse at 50hz
 
#27 ·

Quote:
Originally Posted by Nektarios /forum/post/19383233


So you are saying that they flicker even on low brightness?

The same flicker frequency remains but it is significantly less noticeable due to the smaller amplitude.

Quote:
Originally Posted by Nektarios /forum/post/19383233


Yes I agree about the eye-tracking problem but that is not motion blur, it is judder. You can have judder with a CRT with a refresh rate of 120hz but with a video running at 24 fps. This is because it must redraw each video frame 5 times and that is causing judder to the eye tracking.


LCD Motion blur is caused by the slow response time of the crystals. On an LCD with 120Hz refresh rate (real input refresh rate not the 120hz/240hz interpolation) running a 120 real fps video you see a fading blur/trailing behind the mobing objects because of the slow response time.

It is not judder it is motion blurring. In modern LCD displays (with single digit response times) the primary contribution of motion blur is the hold-type problem (100% duty cycle). This has been ubiquitous throughout literature for many years now. Below is an excellent graphic describing the issue.





Quote:
Originally Posted by Nektarios /forum/post/19383233


Philips on its newest lcd models have real 1ms panels (sharp?) and I have seen it playing high framerate videos with 120hz/240hz interpolation with absolutely no motion blur and no judder, just like a CRT. Besides it there were other LCD models with the same input and with 120hz/240hz interpolation but having motion blur.

I can't comment on this as I have not seen this demo.
 
#28 ·
From what I understand LCD response time need to take with a pinch of salt cause the crystal may respond in say 3ms but they do so in step function and gradation/ grey scale suffer as a result, hence contributing to motion blur. So it is hard to say which problem did Phillips fix since at 3ms you can already refresh at 240Hz mathematically. But LCD definitely need high fps input or interpolation to reduce hold effect blur unlike plasma. It is technically not judder due to source or processing/telecine but the result of retina persistence from hold effect.


I would suggest that the same high framerate input into a theoretical 120 Hz plasma will have only marginal improvement.

Quote:
Originally Posted by specuvestor /forum/post/0


2) generally most people do not see flicker on most content around 50/60Hz but you are right that around 75Hz even extreme contents have no flicker for most people. Almost all cannot see flicker around 85 Hz. That is why plasma telecine pulldown is 72/96Hz which is quite optimal already. 120Hz is a good algorithmic convergence for 24/25/30/60 fps content but not very perceivable or useful for plasma realistically, but useful for LCD as explained above.
 
#29 ·
I tested out a 30p source file on my Panasonic G25 (recorded a game using Fraps), and I also noticed a slight stuttering. Not as bad as 3:2 pulldown, but I still noticed it. Even though 30 divides into 60 which I don't think would need the 3:2 pulldown thing.


I wonder what causes flicker if the source file divides into 60 evenly? If it were the 60 Hz, I would see it everywhere, but I only noticed it during panning. I was told the panning stutters were caused by 3:2 pulldown, but this is a 30p which divides evenly into 60.


Could it have anything to do with my vision? Or is it the TV? Just want to rule some things out.
 
#30 ·

Quote:
Originally Posted by magnetite /forum/post/19388279


I tested out a 30p source file on my Panasonic G25 (recorded a game using Fraps), and I also noticed a slight stuttering. Not as bad as 3:2 pulldown, but I still noticed it. Even though 30 divides into 60 which I don't think would need the 3:2 pulldown thing.


I wonder what causes flicker if the source file divides into 60 evenly? If it were the 60 Hz, I would see it everywhere, but I only noticed it during panning. I was told the panning stutters were caused by 3:2 pulldown, but this is a 30p which divides evenly into 60.


Could it have anything to do with my vision? Or is it the TV? Just want to rule some things out.

It could just be that the pan was too fast. 30p and 24p video will still look "choppier" than 60i/60p video.
 
#31 ·
This is different from the main topic of the thread. This is due to low frame rate which will affect panning for eg 24fps movie. Read in other threads that is why directors have to be careful when doing panning scenes in movies. So it is normal u see stutter during panning. Low frame rate gives an artistic unreal feel to the movie content, which some enjoy.


With ref to the thread, you will see flicker if it were 30fps@30Hz but at 60Hz 2:2 pulldown most of us don't see it.

Quote:
Originally Posted by magnetite /forum/post/0


I tested out a 30p source file on my Panasonic G25 (recorded a game using Fraps), and I also noticed a slight stuttering. Not as bad as 3:2 pulldown, but I still noticed it. Even though 30 divides into 60 which I don't think would need the 3:2 pulldown thing.


I wonder what causes flicker if the source file divides into 60 evenly? If it were the 60 Hz, I would see it everywhere, but I only noticed it during panning. I was told the panning stutters were caused by 3:2 pulldown, but this is a 30p which divides evenly into 60.


Could it have anything to do with my vision? Or is it the TV? Just want to rule some things out.
 
#32 ·

Quote:
Originally Posted by xrox /forum/post/19383361


The same flicker frequency remains but it is significantly less noticeable due to the smaller amplitude.


It is not judder it is motion blurring. In modern LCD displays (with single digit response times) the primary contribution of motion blur is the hold-type problem (100% duty cycle). This has been ubiquitous throughout literature for many years now. Below is an excellent graphic describing the issue.






I can't comment on this as I have not seen this demo.
Attachment 189344


The graph you posted assumes incorrectly that the state changes of the crystals are instantaneous, either on or off whereas in reality the state of the crystals gradually change with time, that's what the response time means and that's what's causing lcd motion blurring.
 
#33 ·

Quote:
Originally Posted by specuvestor /forum/post/19381650


Guys do update us if you still see flicker at 720p60 @60Hz or 1080p24 @72/96Hz.


Nektarios, technically plasma do not leave sufficient retina persistence due to luminence drop for each frame.

Agreed, but you didn't see that I said displaying the same frame multiple times = effective retinal persistence, do you disagree on that?

Quote:
And technically flicker in CRT should not be uniform as luminence output is different across the screen but is perfect for human perception, which form the basis of gamma correction. Not sure how plasma handle the human perception adjustment though.

I don't understand why a scanning gun that uniformly scans the entire screen line by line at the exact same speed, eventually produce different amounts of flicker on different sections of it. It doesn't make sense.
 
#34 ·

Quote:
Originally Posted by Nektarios /forum/post/19388724

Attachment 189344


The graph you posted assumes incorrectly that the state changes of the crystals are instantaneous, either on or off whereas in reality the state of the crystals gradually change with time, that's what the response time means and that's what's causing lcd motion blurring.

I'm pretty sure the author knows what response time is
The graphic is describing the hold-type phenomenon. It is complicated enough without introducing RT.


Do you realize that 120Hz refresh was first considered for LCD to combat "hold-time" blurring?


I do not expect you to change your belief based on my explanations. So hopefully you will consider the large amount of scientific literature on this topic.


Here is a good one that directly addresses the issue of contention:
Hold-Type blurring
 
#35 ·
Nektarios,


If you read that link I posted we could speculate that the Phillips demo (1ms RT) you observed may in fact be true as the hold-type blur issue is attenuated via the 120Hz+interpolation in both displays to the point where the RT again becomes dominant.
 
#36 ·

Quote:
Originally Posted by Nektarios /forum/post/19383233


Philips on its newest lcd models have real 1ms panels (sharp?) and I have seen it playing high framerate videos with 120hz/240hz interpolation with absolutely no motion blur and no judder, just like a CRT. Besides it there were other LCD models with the same input and with 120hz/240hz interpolation but having motion blur.


I just really hope they make a pc monitor out of that panel.

This confuses me. If they can already achieve perfect motion resolution as long as the framerate is high enough (albeit via interpolation), why can't they just put in, say, a blanking frame instead of those interpolated frames, and achieve perfect motion resolution at normal lower framerates. Would it introduce flicker? Some other artefact perhaps?
 
#37 ·
I really don't understand whole motion blur issue... I had a 50/60hz Sony LCD RP-TV and it handled normal TV/dvd/xbox360 gaming better than my current Samsung 63" 3d plasma because of motion blur!!

60fps gaming and 50i TV shows fine on both but when showing "25p" tv material or 30fps gaming it really tends to be less enjoyable with plasma as my old LCD blurs fast movements little bit and just seems smoother. In plasma 25/30fps material just seems jerky/sudden and creates more "double vision" by showing same frame 2 times a row without any buildin blurring (with a bluray it is .... "4 vision", during fast panning I sometimes see 4 copies of same frame 24*4 = 96hz, just eye/brain thing I know..)


Well at least with this plasma 60fps gaming is free of this frame doubling, tripling... issue, I had a Samsung 120hz LCD tv few days and it doubled 60hz to 120hz even during gaming with game mode on!!!! just horrible double vision



So I can not understand why to have any of this 120/240hz lcd stuff with backlight flickering, black frame insertion because plain old 50/60hz LCD tech was just fine perfect, absolutely no flickering and smooth movement with 50/60hz & 25/30hz material.
 
#38 ·

Quote:
Originally Posted by Nektarios /forum/post/19388840


Agreed, but you didn't see that I said displaying the same frame multiple times = effective retinal persistence, do you disagree on that?


I don't understand why a scanning gun that uniformly scans the entire screen line by line at the exact same speed, eventually produce different amounts of flicker on different sections of it. It doesn't make sense.

Generally displaying same frame multiple times does not create retina persistence (I am excluding brain washing here
), guess eyes don't have IR... The persistence is proportionate to luminance and I think contrast. Hence u get retina persistence if shown for eg head lights. In a dark background with single light your retina will also have the after image of the light.


So assuming a plasma is showing a STILL picture at 60fps, you will not get retina persistence, as per xrox chart, the luminance tapers off fast enough for that effect not to happen. But an LCD still picture at 60fps is almost the same as displaying it at 1fps, very sharp and clear. On the other hand if plasma is 1fps, you will see it go black. Hence LCD is perfect for still and PC use. But once you change from still to moving at 60fps, you will notice the blur in the LCD. But fairer to say part of the reason is also the lag in the crystal transition.


The distance from the gun to the bottom and top of the CRT screen vs the middle is different. Hence it is brighter at the middle. But this is perfect as the human eye will perceive them as normal on a 2.2 gamma. There's a lot of literature on this and I must admit I don't understand all of them, hence my previous question if modern plasma actually adjust for this human perception. I am assuming it doesn't because D-Nice has categorically say 2.2 is not the right figure for FPTV calibration.

Quote:
Originally Posted by sonic_blue /forum/post/19389562


This confuses me. If they can already achieve perfect motion resolution as long as the framerate is high enough (albeit via interpolation), why can't they just put in, say, a blanking frame instead of those interpolated frames, and achieve perfect motion resolution at normal lower framerates. Would it introduce flicker? Some other artefact perhaps?

They did. It is called Black Frame Insertion (BFI). BFI in LCD is trying to emulate the luminance drop like the plasma to reduce the hold effect, as explained above, but I think not very successful due to the slow crystal transition. Hence newer LCD uses MCFI rather than BFI.

Quote:
Originally Posted by Jogi /forum/post/19389885


I really don't understand whole motion blur issue... I had a 50/60hz Sony LCD RP-TV and it handled normal TV/dvd/xbox360 gaming better than my current Samsung 63" 3d plasma because of motion blur!!

60fps gaming and 50i TV shows fine on both but when showing "25p" tv material or 30fps gaming it really tends to be less enjoyable with plasma as my old LCD blurs fast movements little bit and just seems smoother. In plasma 25/30fps material just seems jerky/sudden and creates more "double vision" by showing same frame 2 times a row without any buildin blurring (with a bluray it is .... "4 vision", during fast panning I sometimes see 4 copies of same frame 24*4 = 96hz, just eye/brain thing I know..)


Well at least with this plasma 60fps gaming is free of this frame doubling, tripling... issue, I had a Samsung 120hz LCD tv few days and it doubled 60hz to 120hz even during gaming with game mode on!!!! just horrible double vision



So I can not understand why to have any of this 120/240hz lcd stuff with backlight flickering, black frame insertion because plain old 50/60hz LCD tech was just fine perfect, absolutely no flickering and smooth movement with 50/60hz & 25/30hz material.

Frankly Jogi, I am not sure I understand your post. Firstly LCD RPTV and LCD FPTV are very different unless I am missing something here? Secondly you are the only one I know that can see motion frame by frame... if that is true I have to withdraw my skepticism towards
 
#39 ·

Quote:
Originally Posted by Jogi /forum/post/19389885


I really don't understand whole motion blur issue...

It's quite simple really, LCD's blur under motion and it's not supposed to look that way. By the sounds of it you seem to prefer the blurry motion due to it masking jitter on lower framerates like 24/25/30fps. I don't understand how anyone can complain about the motion jitter of low framerates as that is how they've always appeared on CRT's and is how they are supposed to appear. It does not look bad at all. I guess people are all swept up in the hyper smooth soap opera motion that they now want artificially smooth images. Quite sad really...
 
#40 ·

Quote:
Originally Posted by Jogi /forum/post/19389885


I really don't understand whole motion blur issue... I had a 50/60hz Sony LCD RP-TV and it handled normal TV/dvd/xbox360 gaming better than my current Samsung 63" 3d plasma because of motion blur!!

60fps gaming and 50i TV shows fine on both but when showing "25p" tv material or 30fps gaming it really tends to be less enjoyable with plasma as my old LCD blurs fast movements little bit and just seems smoother. In plasma 25/30fps material just seems jerky/sudden and creates more "double vision" by showing same frame 2 times a row without any buildin blurring (with a bluray it is .... "4 vision", during fast panning I sometimes see 4 copies of same frame 24*4 = 96hz, just eye/brain thing I know..)

This "double vision" you are talking about, is called Judder and is what I am talking about. On 60fps (or the same fps as your refresh rate is set to) you should not see any blurring/judder on a CRT and plasma (most plasmas leave phosphor trails though).


On the other hand, the "blurring", that paradoxically you found it more pleasing, is caused by the lcds slow RT because it cannot transit from one frame to the other "instantaneous", but instead it gradually shifts the colors and thereby causing a slight blur.
 
#41 ·

Quote:
Originally Posted by specuvestor /forum/post/19389986


They did. It is called Black Frame Insertion (BFI). BFI is trying to emulate the luminance drop like the plasma, as explained above, but I thin not very successful due to the slow crystal transition. I think newer LCD uses MCFI rather than BFI.

You're right, if the LCD motion blur was caused of a 100% duty cycle then 120hz LCDs or BFI would totally eliminate it, but they don't because then the slow RT blurring still exists.
 
#43 ·

Quote:
Originally Posted by sonic_blue /forum/post/19390002


Another thing to note is that low framerate rendered computer graphics do not have motion blur which occurs in cameras as a result of slow shutter speed.


So a 30fps video game is going to look a lot more jittery than 30fps from a video camera.

Yes that's correct! That's why even at 24fps movies are watchable enough. However there are some games (crysis comes to mind) that do artificially motion blurring and it does mitigate the problem. But that's not really a solution because the human vision doesn't like motion blur when it needs to track a moving image, the motion must be smooth yet very clear not blurry.
 
#44 ·

Quote:
Originally Posted by xrox /forum/post/19388971


I'm pretty sure the author knows what response time is
The graphic is describing the hold-type phenomenon. It is complicated enough without introducing RT.


Do you realize that 120Hz refresh was first considered for LCD to combat "hold-time" blurring?


I do not expect you to change your belief based on my explanations. So hopefully you will consider the large amount of scientific literature on this topic.


Here is a good one that directly addresses the issue of contention:
Hold-Type blurring

I am sorry but it is clear that the author, knowing what it is or not, he doesn't take into account the RT in that graph.


I haven't changed my belief about it, but I have to thank you and everybody here for all the information and your contribution because we all gain from it.

Quote:
Originally Posted by xrox /forum/post/19389019


Nektarios,


If you read that link I posted we could speculate that the Phillips demo (1ms RT) you observed may in fact be true as the hold-type blur issue is attenuated via the 120Hz+interpolation in both displays to the point where the RT again becomes dominant.
Quote:
LCD Motion blur is a well known problem. Although many solutions have been proposed, some fundamental questions have not been answered yet. In this paper, we try to answer such questions. Specifically, we calculate the waveform and its blur width of a moving edge perceived on LCD screen for current LCD and the proposed four solutions of hold-type motion blur. We found that the slow response of current LCD is not a dominant factor of motion blur. The slow response of current LCD only contributes to 30% of the motion blur, while the hold-type rendering mode of LCD contributes to 70%. Therefore, fast LCD such as OCB itself does not significantly reduce motion blur. Fast LCD, on the other hand, is critical to the proposed three solutions of hold-type blur to avoid the ghosting artifact. With fast LCD, black data insertion and frame rate doubling can provide 50% reduction of motion blur. With both fast response LCD and fast backlight, backlight flashing can provide much higher reduction of motion blur.

According to the scientific paper you referred me to (which I note that it is from 2005, when LCDs had 60hz refresh rates, compared to 120hz of today's), it states clearly that 70% of blurring is caused by the hold-type rendering and 30% is caused by the slow RT. That's exactly what I am supporting here.


When you go up to 120hz refresh rate, the hold type problem is effectively solved, but the slow RT blur still remains.


Yes the philips TV must have both high refresh rate and really very fast response time, or I can't see any other explanation. The novelty of it, is the really fast RT, because high refresh rates exist for years now. I am confident that this tech has solved the motion blur problem with LCDs and I really want to see it in a PC monitor!
 
#45 ·

Quote:
Originally Posted by Nektarios /forum/post/19390033


According to the scientific paper you referred me to (which I note that it is from 2005, when LCDs had 60hz refresh rates, compared to 120hz of today's), it states clearly that 70% of blurring is caused by the hold-type rendering and 30% is caused by the slow RT. That's exactly what I am supporting here.


When you go up to 120hz refresh rate, the hold type problem is effectively solved, but the slow RT blur still remains.

Yes... IF (Capital letters here) they are fed 120fps input or MCFI, as we cannot perceive difference above 85Hz. That's why MCFI is important in LCD despite the soap opera effect but quite useless in plasma, as per my answer to the OP and the purpose of this thread.


I am skeptical about LCD refresh rate as it doesn't have a real standard due to the different response time. So even if a 8ms LCD can do mathematically 125Hz refresh, we will still see blur not because the refresh is not high enough, but because of the slow RT leaving image trail. It does complicate matters a bit
Hence the marketing jargons on 240Hz and 600Hz etc.


"Good thing" only a "selected" few has problems with Phosphor trail.
 
#46 ·

Quote:
Originally Posted by Nektarios /forum/post/19390033


According to the scientific paper you referred me to (which I note that it is from 2005, when LCDs had 60hz refresh rates, compared to 120hz of today's), it states clearly that 70% of blurring is caused by the hold-type rendering and 30% is caused by the slow RT. That's exactly what I am supporting here.

Actually from your earlier posts you did not support this at all. In fact you did not even know what hold-type blur was. But now you do. But as I clearly stated in my earlier post I can only "speculate" that is what is happening in the Phillips demo.


As for hold-type being solved by 120Hz MCFI or BFI I'm not convinced. The hold time would be halved to 8ms which is still higher than plasma (4-6ms) and still much higher than CRT (~1ms).


Also, if date holds value to you I could post more recent literature if you like. That paper states 30% for RT based on RT from 5 years ago. You could use the same argument that todays modern LCD have improved RT from 5 years ago.
 
#47 ·

Quote:
Originally Posted by specuvestor /forum/post/19389986


So assuming a plasma is showing a STILL picture at 60fps, you will not get retina persistence, as per xrox chart, the luminance tapers off fast enough for that effect not to happen. But an LCD still picture at 60fps is almost the same as displaying it at 1fps, very sharp and clear. On the other hand if plasma is 1fps, you will see it go black. Hence LCD is perfect for still and PC use. But once you change from still to moving at 60fps, you will notice the blur in the LCD. But fairer to say part of the reason is also the lag in the crystal transition.

You may have misinterpreted the graphic (I'm not sure). Your eye has to be "moving" to create blur due to retinal persistence. This is why they show an eyeball in movement. Essentially what the graphic is saying is that the longer each frame is displayed (either via duty cycle (left chart) or repeated frames + duty cycle (right chart)) the greater the blur on your "MOVING" retina that is tracking movement on the display.


Actually, I think you understand OK.
 
#48 ·

Quote:
Originally Posted by specuvestor /forum/post/19389986


They did. It is called Black Frame Insertion (BFI). BFI in LCD is trying to emulate the luminance drop like the plasma to reduce the hold effect, as explained above, but I think not very successful due to the slow crystal transition. Hence newer LCD uses MCFI rather than BFI.

BFI is not popular because it is too limiting. Not only does it reduce the brightness which is LCD strongpoint but it does not enable judder reduction as well as MCFI. Furthermore when displaying film MCFI enables greater reduction of hold time than BFI.


What RT does limit is backlight scanning due to the ghosting effect. The backlight has to be synced perfectly with the pixel array scan.
 
#49 ·

Quote:
Originally Posted by xrox /forum/post/19390291


Actually from your earlier posts you did not support this at all. In fact you did not even know what hold-type blur was. But now you do. But as I clearly stated in my earlier post I can only "speculate" that is what is happening in the Phillips demo.


As for hold-type being solved by 120Hz MCFI or BFI I'm not convinced. The hold time would be halved to 8ms which is still higher than plasma (4-6ms) and still much higher than CRT (~1ms).


Also, if date holds value to you I could post more recent literature if you like. That paper states 30% for RT based on RT from 5 years ago. You could use the same argument that todays modern LCD have improved RT from 5 years ago.

Yes I did support it, it's right there and I know what hold-type problem was for years now. What I disagree with you is that you say that low fps/retinal persistence blur/trailing/judder is the same as crystal RT blur/trailing/ghosting, which is not.


What I add is that hold-type is practically solved with high refresh rates but RT and especially RTC Artifacts that get as high as 25ms+ remains the problem now.
 
#50 ·

Quote:
Originally Posted by Nektarios /forum/post/19390538


Yes I did support it, it's right there and I know what hold-type problem was for years now. What I disagree with you is that you say that low fps/retinal persistence blur/trailing/judder is the same as crystal RT blur/trailing/ghosting, which is not.

Sorry but from your earlier posts it did seem you did not understand it. Maybe you just mispoke?

Quote:
Originally Posted by Nektarios /forum/post/19390538


Yes I agree about the eye-tracking problem but that is not motion blur, it is judder.
 
#51 ·

Quote:
Originally Posted by Nektarios /forum/post/19390538


What I add is that hold-type is practically solved with high refresh rates but RT and especially

I don't agree. If you list hold-times of various display tech and then list them in order of reputation for blur the list will be sorted in relative to hold time.


Remember I'm listing "hold-time" here and not RT


PMOLED - microseconds

CRT - 1 ms

PDP - 4-6ms

120Hz LCD (MCFI) - 8.3ms

AMOLED - 16.7 ms

60Hz LCD - 16.7ms


Now as I speculated, RT may be more important when hold time is reduced to 8ms. Or maybe not. I have a paper from 2009 that suggests that RT blur is negilgible with modern displays with RT in the single digits.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top